dns-operator-route53 is a controller, which runs once per management Cluster. It's responsible for creating the following DNS Records per cluster:
A: api.<clustername>.test.gigantic.io (points to the kubernetes API IP of a cluster)
A: bastion1.<clustername>.test.gigantic.io (points to the bastion Host IP of a cluster - only on OpenStack yet)
A: ingress.<clustername>.test.gigantic.io (points to kube-system/nginx-ingress-controller)
CNAME: *.<clustername>.test.gigantic.io for ingress.<clustername>.test.gigantic.io
reconciliation loop
Even there is no need at all to set the finalizer dns-operator-route53.finalizers.giantswarm.io on some infrastructureProviders but to keep the
behavior as equal as possible over different infrastructureProviders we do so.
new infrastructure provider
extend RBAC by adapting infraCluster function in _helpers.tpl file.
if the new infrastructureProvider has the concept of a bastion host, extend the BastionIP method of ClusterScope to extract.
known issues
better error handling during cluster creation/deletion
we currently get a bunch of ERRORs during cluster creation/deletion which we didn't catch up correctly. Due the reconciliation it took some iterations once we don't get any new errors on a cluster base.